The Hierarchical Structure of Childhood Personality
in Five Countries: Continuity From Early Childhood
to Early Adolescence
Jennifer L. Tackett,1 Helena R. Slobodskaya,2
Raymond A. Mar,3 James Deal,4
Charles F. Halverson, Jr.,5 Spencer R. Baker,6
Vassilis Pavlopoulos,7 and Elias Besevegis7
1
2
University of Houston
Siberian Branch of the Russian Academy of Medical Sciences
3
4
York University
North Dakota State University
5
University of Georgia
6
Hampton University
7
University of Athens
ABSTRACT Childhood personality is a rapidly growing area of investigation within individual differences research. One understudied topic is
the universality of the hierarchical structure of childhood personality. In
the present investigation, parents rated the personality characteristics of
3,751 children from 5 countries and 4 age groups. The hierarchical structure of childhood personality was examined for 1-, 2-, 3-, 4-, and 5-factor
models across country (Canada, China, Greece, Russia, and the United
States) and age group (3–5, 6–8, 9–11, and 12–14 years of age). Many
similarities were noted across both country and age. The Five-Factor
Model was salient beginning in early childhood (ages 3–5). Deviations
This research was partially supported by a grant from the Social Sciences and Humanities Research Council of Canada to JT and Russian Humanitarian Foundation grant
#11-06-00517a to HS. The authors would like to thank the many families who participated in this research and the researchers who facilitated data collection.
Correspondence concerning this article should be addressed to Jennifer L.
Tackett, Department of Psychology, University of Houston, Houston, TX, 77204.
Email: jtackett@uh.edu.
Journal of Personality 80:4, August 2012
© 2011 The Authors
Journal of Personality © 2011, Wiley Periodicals, Inc.
DOI: 10.1111/j.1467-6494.2011.00748.x
2
Tackett, Slobodskaya, Mar, et al.
across groups and from adult findings are noted, including the prominent
role of antagonism in childhood personality and the high covariation
between Conscientiousness and intellect. Future directions, including the
need for more explicit attempts to merge temperament and personality
models, are discussed.
Individual differences in personality have primarily been studied
with adult populations, partly due to the availability and administrative ease of existing adult personality measures. Early attempts
to apply adult personality measures to children and adolescents
yielded promising results, however, suggesting that self- and informant report questionnaires could be applied to younger age groups
(Tackett, 2006). This work, in combination with the extensive literature on temperament, offered a substantial foundation for seminal
reviews on the conceptualization and measurement of childhood
personality (Caspi, Roberts, & Shiner, 2005; Shiner & Caspi, 2003).
As the burgeoning field of childhood personality research continues to grow, advances have been made in creating developmentally appropriate measures (e.g., Barbaranelli, Caprara, Rabasca,
& Pastorelli, 2003; Halverson et al., 2003; John, Caspi, Robins,
Moffitt, & Stouthamer-Loeber, 1994; Mervielde & De Fruyt, 1999).
With improved measurement comes better research on outcomes,
and childhood personality is now recognized as a demonstrated predictor of many important adult outcomes (e.g., physical and mental
health, adaptive relationships; Hampson, 2008).
Further research is needed to capitalize on these promising early
successes and tackle existing gaps in our understanding of childhood
personality. The present study focuses on one such unstudied issue:
the nature of hierarchical personality structure in younger populations. Hierarchical structures are targeted for investigation in the
present study because they hold both theoretical (e.g., understanding
the development of trait relationships) and practical (e.g., facilitating
integration of results across studies) implications for this area of
research, representing a necessary feature for understanding personality variation (Markon, 2009).
Childhood Personality Across Development
Multiple studies have examined mean-level change and differential
stability of personality traits prior to adulthood (e.g., Allik, Laidra,
Realo, & Pullmann, 2004; De Fruyt et al., 2006; McCrae et al., 2002;
Child Personality Structure
3
Roberts & DelVecchio, 2000; Roberts, Walton, & Viechtbauer,
2006). Most of this research has investigated personality traits within
the Five-Factor Model (FFM), the predominant approach to measuring personality in adults (John, Naumann, & Soto, 2008). The
FFM includes five higher order domains: Neuroticism (N), Extraversion (E), Agreeableness (A), Conscientiousness (C), and Openness to
Experience (O). This structure, while well established within adults,
has not been as thoroughly examined within younger populations.
Merging this FFM literature with research on temperament will
lead to a better understanding of personality structure across childhood. In order to do so, it is necessary to reconcile potential differences in temperament and personality models, some of which parallel
previous debates in adult personality research. For example, Rothbart’s prominent temperament measures converge on three, rather
than five, higher order factors: Extraversion/Surgency, Negative
Affectivity, and Effortful Control (e.g., Rothbart, Ahadi, Hershey, &
Fisher, 2001). Similar to three-factor models in adults (e.g., Tellegen
& Waller, 1992), two of the FFM traits are represented in this model:
E and N (although N often includes components of disagreeableness
as well). The third trait reflects a broader disinhibitory construct,
likely reflecting both A and C components (Caspi et al., 2005). Thus,
in temperament models, A does not emerge as a higher order trait
early in development, although an analogous trait (affiliativeness) is
present in adolescent/adult measures (Evans & Rothbart, 2009).
Further, in both temperament and adult three-factor models, O as a
higher order domain is missing. In models of childhood personality,
A is sometimes represented by aspects of agreeable compliance and
antagonism more so than the typical FFM models in adults (De Pauw
& Mervielde, 2010; Digman & Shmelyov, 1996; Goldberg, 2001;
Tackett, Krueger, Iacono, & McGue, 2008). Openness to Experience
is often marked by curiosity, intellect, and imagination in childhood
(Gjerde & Cardilla, 2009; Goldberg, 2001; Halverson et al., 2003)
without including items such as cultural interests and aesthetic appreciation, items that are more common in adult measures. Taken
together, these findings suggest that early models of individual differences often do not include A and O and, when they do, they are
commonly measured somewhat differently than in adults.
It has been hypothesized that early temperamental traits
break down into more complex personality traits across development
(e.g., Digman & Shmelyov, 1996; Rothbart, Ahadi, & Evans, 2000;
4
Tackett, Slobodskaya, Mar, et al.
Rothbart & Bates, 2006; Shiner & Caspi, 2003), suggesting that these
taxonomic differences across age may be suggestive of developmental phenomena (Caspi et al., 2005). For example, early life taxonomies suggest that A and C reflect a common trait in childhood. If this
is the case, then an important question is at what point across development do A and C become distinct phenomena? Self-reports of
personality demonstrate that A and C become increasingly differentiated between 10 and 20 years of age (Soto, John, Gosling, & Potter,
2008), but a greater understanding of this potential differentiation
even earlier in life is lacking. A similar question might be asked of O:
When does this trait emerge in development? Previous research has
suggested that O may not be an emergent trait in childhood (Eder,
1990), manifesting only later in adolescence or early adulthood. Such
an explanation would account for the absence of O in early age
temperament models. More recent studies, however, have suggested
that O is salient and measurable by early childhood (De Pauw,
Mervielde, & Van Leeuwen, 2009; Gjerde & Cardilla, 2009). Additional evidence speaking toward these divergent accounts is an
important goal for research in childhood personality.
A primary barrier to addressing these questions has been a lack
of measurement tools that span the relevant ages (De Pauw &
Mervielde, 2010). Relying on adult instruments may not be the ideal
approach to providing such evidence because personality traits may
show heterotypic continuity across development. That is, manifestations of the same underlying trait may look different depending on
the developmental period (Caspi et al., 2005; De Fruyt, Mervielde,
Hoekstra, & Rolland, 2000). Difficulty extracting A or O in a childhood sample may be the result of measurement limitations rather
than evidence against the presence of A and O as distinct traits at
early ages (De Pauw & Mervielde, 2010). Examination of different
levels of the personality hierarchy across age, using age-appropriate
measures, will offer important insight into these issues. Discussion of
personality structure also requires consideration of similarities and
differences across culture, in light of the interest in personality as a
potential human universal (Lee & Ashton, 2008).
Childhood Personality Across Cultures
Just as understanding personality structure across age aids our development of comprehensive trait models, so does cross-cultural replication (Bornstein, 2002; Caspi et al., 2005). In recognition of this
Child Personality Structure
5
fact, Caspi and colleagues (2005) described cross-cultural generalizability of personality structure as one of the most important future
directions in childhood personality research. Evidence for crosscultural similarities (and differences) in personality traits not only
offers information in terms of mean-level distinctions but also helps
to reveal underlying processes reflected in manifest traits (Rothbart
et al., 2000). Thus, cross-cultural research in childhood personality
offers an important avenue for better understanding convergence
and divergence at the phenotypic level as well as the etiologic level.
There are many important measurement considerations when comparing personality across cultures (Church, 2001), and rigorous
cross-cultural replications are difficult to achieve. Cross-cultural personality research on adults has generally proven robust replication of
the FFM, with respect to content and predictive validity (Church,
2001). However, of the factors in the FFM, O has generally received
the weakest support across cultures (Church & Lonner, 1998; Heine
& Buchtel, 2009). This similarity across cultures provides further
support to the universality of trait structure, an important feature
when investigating applicability to younger ages. Establishing robust
trait structure across cultures at younger ages would support valid
measurement of analogous traits at younger ages.
Culture is thought to be one of the most important environmental
influences on personality development (Rothbart & Bates, 2006;
Super & Harkness, 2002). Most previous studies of child personality
have been conducted in Western cultures, typically involving participants from only one country, limiting the generalizability of these
findings to younger populations as a whole (Gartstein, Peleg, Young,
& Slobodskaya, 2009). Some cross-cultural studies of childhood personality have focused on examination of mean differences (e.g.,
Knyazev, Zupancic, & Slobodskaya, 2008), revealing some robust
differences across individualistic and collectivistic cultures. For
example, children in collectivistic cultures (e.g., China, Japan, and
Korea) are often rated as more behaviorally inhibited than those
in individualistic cultures (e.g., Australia, Canada, Italy, and the
United States; Ahadi, Rothbart, & Ye, 1993; Chen et al., 1998;
Rubin et al., 2006; Windle, Iwawaki, & Lerner, 1988). Limited
studies investigating emergent factor structures across cultures have
supported measurement equivalence for existing measures at early
ages (e.g., Ahadi et al., 1993; Windle et al., 1988). To the best of our
knowledge, a cross-cultural investigation of the hierarchical struc-
6
Tackett, Slobodskaya, Mar, et al.
ture of childhood personality has not yet been undertaken. This is a
particularly important issue, as trait hierarchy offers a more nuanced
and flexible approach to examining group differences in the causal
processes and behavioral manifestations of traits (Markon, 2009).
One of the primary strengths of the current study is the employment of the same measure of childhood personality across cultures.
Moreover, an advantage of this measure is the developmental origins
of the item pool, which emerged from a large and comprehensive
cross-cultural approach to developing a taxonomy of childhood personality (Halverson et al., 2003; Kohnstamm, Halverson, Mervielde,
& Havill, 1998). Although we draw upon cross-cultural work to
inform the present study, it is important to note that the concept of
culture is more complex than country of origin (Kotelnikova &
Tackett, 2010; Schwartz, 1992). With that caveat, many cultural
frameworks map onto specific countries, allowing for some generalization. Specifically, this investigation uses participants from countries that would typically be classified as collectivistic (i.e., China)
and individualistic (i.e., Canada and the United States; Hofstede,
2001; Suh, Diener, Oishi, & Triandis, 1998). Of the other countries
investigated, both are scored as more similar to collectivistic countries (Greece and Russia). We use the term country rather than
culture in the present study to clearly indicate that a nuanced
measure of culture was not utilized in this investigation.
Advances in Structural Models of Personality
Hierarchical structural models of personality traits have gained the
attention of personality psychologists with the publication of Digman’s (1997) important work on the existence of broad traits above
the Big Five. The utility of hierarchical models has proven to be a
useful tool in reconciling previous debates regarding differences
and similarities among factorial structures. Markon, Krueger, and
Watson (2005) utilized a hierarchical structural approach to demonstrate empirical connections among two-, three-, four-, and fivefactor models of personality in adults. However, such work in adults
has not been entirely conclusive, as similar investigations with adult
temperament traits reveal a slightly different pattern of association.
Specifically, A tends to merge with E at higher levels rather than with
N and C (Evans & Rothbart, 2009).
Hierarchy has been described as “an intrinsic and pervasive
feature of trait structure” (Markon, 2009, p. 812). Thus, a compre-
Child Personality Structure
7
hensive understanding of trait structure in early life must incorporate
investigations of hierarchy. Understanding trait hierarchies holds
numerous advantages, both theoretical and practical (Markon,
2009). Hierarchies allow much greater flexibility in understanding
trait structure than do simplistic models (which are more commonly
evoked) and thus allow more nuanced explanations for causal theories of traits. This is particularly important when examining group
differences, including differences across age or country, as we investigate in the present study. Practically speaking, trait hierarchies
provide integrative structural frameworks within which various
theoretical models and empirical results can be interpreted. This will
facilitate further integration of the rich areas of temperament and
child personality research.
Goldberg (2001) examined Digman’s teacher ratings of childhood
personality for children in Grades 1–6 via factor analysis and found
support for the FFM. He further analyzed factor intercorrelations
among the FFM as potential evidence of Digman’s two-factor structure. He found broad support for a two-factor structure resembling
that found with adults with 1 deviation, in that C and intellect were
more highly correlated in children than was typically demonstrated
in adults. In later work, longitudinal trait relationships for this
sample in adulthood were more robust at this broader two-factor
level than at the five-factor level (Hampson & Goldberg, 2006).
Recent investigations (Martel, Nigg, & Lucas, 2008; Tackett et al.,
2008) of the hierarchical structure of personality and temperament in
childhood (children ranging in age from 7 to 13) have roughly paralleled the results established by Markon et al. (2005). These results
show strongest convergence with the adult findings at the two-factor
level, also consistent with Goldberg’s (2001) investigation. Deviations from adult findings were also noted. In both studies, negative
emotionality covaried highly with aspects of (dis)agreeableness
(Martel et al., 2008; Tackett et al., 2008), positioning antagonistic
traits in a more prominent role than is typically found in adult
models. Numerous questions similarly remained for the role of O or
intellect in these structures. Neither previous investigation included
close analogs to O or intellect (Martel et al., 2008; Tackett et al.,
2008), although one study included items measuring absorption
(Tackett et al., 2008). Further, previous studies have not yet examined the full developmental range from early childhood to early
adolescence. Thus, we build on this previous work with a more
8
Tackett, Slobodskaya, Mar, et al.
comprehensive personality assessment and the ability to examine
differences across age groups and countries.
The Present Study
In the present study, we sought to reconcile some remaining questions in this area of study by examining the hierarchical structure of
childhood personality. In a large sample of children ranging in age
from 3 to 14 years old and originating from five different countries
around the world, we conducted a series of analyses examining itemlevel covariation of parent-reported personality characteristics. We
first compared emergent hierarchical structures across five countries:
Canada, China, Greece, Russia, and the United States. We next
compared emergent hierarchical structures across four age groups:
early childhood (3–5-year-olds), middle childhood (6–8-year-olds),
later childhood (9–11-year-olds), and early adolescence (12–14-yearolds). The cross-country analyses were largely exploratory, given the
lack of relevant work on cultural differences in childhood personality. Based on adult work, we expected to find strong similarities
across countries for E, N, A, and C, with less robust emergence of O
across countries. In the cross-age analyses, we hypothesized that
evidence for two- and three-factor models would remain robust
across development, with evidence for four- and five-factor models
appearing more clearly at later ages.
METHOD
Participants
Canada. The Canadian sample included caregiver report data for 392
children (49.7% female) from a large metropolitan area in southern
Ontario. Caregivers were primarily mothers (95.9%), with the remainder
being fathers (4.1%). Data from other caregivers were used in the present
analyses only when mother report was not available, and this is true for
the remaining samples as well. Mean age for the target children was 9.31
years (SD = 1.91), ranging from 4 to 13 years of age. Participants were
recruited through a variety of methods, including use of a database maintained by the Department of Psychology consisting of interested parents,
flyers, newspaper advertisements, and via the Institute for Child Study,
affiliated with the University of Toronto. Personality data were collected
as part of other larger studies.
Child Personality Structure
9
China. The Chinese sample included caregiver report data for 506 children (50.6% female) from the Dalian region of China. Caregivers were
primarily mothers (98.8%) but also included other caregivers if mother
report was not available (1.2%). Mean age for the target children was 6.72
years (SD = 2.32), ranging from 4 to 11 years of age. Participants were
recruited primarily from schools by faculty and students of the Liao Ning
Teacher’s College in Dalian.
Greece. The Greek sample included caregiver report data for 572
children (48.3% female) from the suburban Athens area. Caregivers were
primarily mothers (96.5%) but also included fathers (3.2%) and other
caregivers if mother report was not available (0.3%). Mean age for the
target children was 10.02 years (SD = 2.45), ranging from 4 to 14 years of
age. Participants were recruited using a snowball sampling method in
which undergraduate and graduate psychology students nominated at
least five families they knew with a preschool- or school-age child. These
families were then contacted by phone. If they agreed to participate in
the study, a research assistant visited them at their home and administered
the questionnaires. These families were then asked to nominate another
family with a preschool- or school-age child.
Russia. The Russian sample included caregiver report data for 1,374
children (47.6% female). Most data (70.5%) were collected in Novosibirsk,
Russia’s third largest city, with the remainder coming from rural Siberian
regions (29.5%). Caregivers were primarily mothers (87.3%) but also
included fathers (8.0%) and other caregivers (4.7%). Mean age for the
target children was 8.63 years (SD = 3.37), ranging from 2 to 14 years of
age. Participants were recruited through a variety of methods: Caregivers
were approached via child care centers, schools, and in person, including
home visits.
United States. The American sample included caregiver report data for
907 children (53.0% female) from Georgia, Virginia, and North Dakota.
Caregivers were primarily mothers (63.2%) but also included fathers
(27.3%) and other caregivers if mother report was not available (6.0%) and
participants for whom caregiver status was not collected (3.5%). Mean age
for the target children was 7.35 years (SD = 3.16), ranging from 2 to 14
years of age. Participants were recruited by undergraduate students who
collected data as a class assignment or extra credit. Students identified and
contacted parents with a preschool- or school-age child; if the parents
agreed to participate, the students provided the parents with a questionnaire packet, which they completed and returned to the investigators.
10
Tackett, Slobodskaya, Mar, et al.
Materials and Procedure
All samples were administered the Inventory for Child Individual Differences (ICID; Halverson et al., 2003), a 144-item measure designed to
assess childhood personality. Items are rated on a 7-point Likert scale,
ranging from 1 (much less than the average child or not at all) to 7 (much
more than the average child). Only 108 of the ICID items are needed to
score the 15 lower order scales, and some data collection in the current
samples utilized this abbreviated version. In the present study, we use
those 108 items that were collected for all participating children to ensure
uniformity across samples. Translations of the Chinese and Greek ICID
measures are described extensively elsewhere (Halverson et al., 2003), as is
validation of the Russian version of the ICID (Knyazev & Slobodskaya,
2005).
Participating caregivers completed the 108- or 144-item ICID
about their child. Caregivers completed the measures at home (Canada,
China, Greece, Russia, and the United States) or in the lab (Canada).
Canadian caregivers received modest monetary compensation for their
participation. In some cases, Russian participants received monetary
compensation or a small gift. The remaining Russian participants and
all participants from the United States, Greece, and China were not
compensated. All of the studies included in this analysis were approved
by the relevant on-site ethics review board.
RESULTS
Scale scores at the five-factor domain level of the ICID items were
calculated for descriptive purposes according to Halverson et al.
(2003). Means and standard deviations of the scored ICID domains
(on a 7-point scale) are presented in Table 1 by country and age
group. Given that the primary goal of the study was to examine
differences in emergent hierarchical structures, we used items in
the following analyses rather than scales that presuppose a given
personality structure in the data.
Goldberg’s “Bass-ackwards” Method (2006)
The hierarchical structure of the 108 ICID items was extracted using
the procedures recommended by Goldberg (2006) for examining hierarchical covariation among a set of items or behaviors.
Adhering to these recommendations, we extracted principal components with varimax rotation at Levels 1–5 of the hierarchy.
United States
4.54 (.79)
3.74 (.67)
5.29 (.65)
4.78 (.73)
5.23 (.76)
Russia
4.50 (.91) 4.49 (.77) 4.69 (.94) 4.32 (.74)
3.68 (.58)
4.95 (.60)
4.70 (.57)
5.00 (.70)
Greece
3.51 (.68)
5.12 (.65)
4.78 (.70)
5.07 (.77)
3.55 (.66)
5.08 (.63)
4.86 (.76)
5.16 (.78)
Neuroticism
Extraversion
Agreeableness
Openness to
Experience
Conscientiousness
China
3.61 (.60)
4.65 (.61)
4.76 (.63)
4.57 (.68)
Canada
Domain Scales
Country
3.66 (.62)
4.92 (.64)
4.72 (.65)
4.86 (.73)
6–8
3.61 (.67)
4.93 (.69)
4.80 (.72)
4.92 (.81)
9–11
3.58 (.68)
4.90 (.73)
4.85 (.71)
4.87 (.86)
12–14
4.51 (.67) 4.40 (.77) 4.46 (.87) 4.54 (.95)
3.58 (.58)
5.03 (.63)
4.72 (.59)
4.98 (.70)
3–5
Age Groups
Table 1
Means and Standard Deviations of Parental Ratings of Childhood Personality Across Countries and Age
Child Personality Structure
11
12
Tackett, Slobodskaya, Mar, et al.
Specifically, one principal component was extracted from all items to
represent Level 1 of the hierarchy in each group. Next, two principal
components were extracted from all items to represent Level 2, and
so on. Regression-based factor scores were saved at each level and
later correlated to provide “path estimates” between contiguous
levels of the hierarchy, allowing for an examination of how higher
levels deconstruct into lower levels. We were primarily interested in
the higher order hierarchical structure of childhood personality;
thus, results are presented for Levels 1–5 from each analysis, with the
top six items defining each component presented (see Figures 1–5).
All analyses were conducted with SPSS.
Personality Hierarchy Across Countries
A comparison of structures at Level 5 from all countries is presented
in Figure 1. Due to space constraints, all five country-specific hierarchies are not represented here, but are available from the first
author upon request. Detailed descriptions of these findings are provided below.
Canada. In each of the following analyses, the single principal
component at Level 1 is extracted from all available items. Level 2 of
the personality hierarchy for Canadian children indicated two broad
components reflecting characteristics of E/O and A/N. Both of these
components contributed to a third component reflecting C/O at
Level 3, with E/O contributing more variance. At Level 4, E breaks
down into E and A. Finally, at Level 5, C/O breaks down into
separate C and O components. The final five-factor structure for
Canadian children does not contain a “pure” N but rather a component primarily defined by antagonism and disagreeableness as
well as a more typical A component. Variance accounted for at
Levels 1–5 was 24.25%, 33.63%, 40.99%, 45.14%, and 48.57%,
respectively.
China. Level 2 of the personality hierarchy for Chinese children
indicated a typical E component and an N/C component that also
included the physical activity characteristics that would typically be
expected to load on E. At Level 3, both components contribute to
new C/O and E components, with only the Level 2 N/C component
contributing to an antagonism-heavy Level 3 A/N factor. At Level 4,
13
Child Personality Structure
Canadian
(n = 392)
Chinese
(n = 506)
Greek
(n = 572)
Russian
(n = 1,374)
American
(n = 907)
Component 1
E
Energetic
Outgoing
Loves to be w/people
Sociable
Lively/enthusiastic
Makes friends easily
C/O
Good concentration
Large vocabulary
Careful
Short attention span (rc)
Quick to learn
Unimaginative (rc)
E
Sociable
Loves to be w/people
Makes friends easily
Cheerful
Lively/enthusiastic
Energetic
E
Energetic
Sociable
Makes friends easily
Physically active
Always on the move
Lively/enthusiastic
E
Outgoing
Loves to be w/people
Energetic
Friendly
Happy
Makes friends easily
Component 2
A/N
Irritable
Angry easily
Quick tempered
Complains
Moody
Disrespectful
A
Loving
Thoughtful of others
Considerate
Helpful
Quick to understand
Caring
A/N
Quick tempered
Stubborn
Hardheaded
Wants things own way
Rude
Angry easily
A/N
Irritable
Quick tempered
Aggressive
Rude
Whiny
Angry easily
A/N
Rude
Selfish
Mean
Quick tempered
Aggressive
Angry easily
Component 3
O
Thinking abilities
Quick to learn
Intelligent
Large vocabulary
Quick to understand
Speaks well
E
Outgoing
Withdrawn (rc)
Lively/enthusiastic
Loves to be w/people
Talkative
Makes friends
C
Organized
Careful
Neat and tidy
Responsible
Perfectionist
Drive to do better
A
Sensitive
Loving
Caring
Thoughtful
Joy to be with
Helpful
O
Quick to learn
Large vocabulary
Good thinking abilities
Intelligent
Good memory
Quick to understand
Component 4
A
Thoughtful
Sensitive
Caring
Loving
Considerate
Sweet
N
Easily upset
Quick tempered
Insecure
Irritable
Feelings hurt
Afraid of lots
O
Quick to learn
Quick to understand
Thinking abilities
Good memory
Intelligent
Slow to learn (rc)
C
Self-disciplined
Good concentration
Organized
Careful
Responsible
Neat and tidy
N
Lacks confidence
Afraid of lots
Feelings hurt
Fearful
Difficulty adjusting
Needs help w/lots
Component 5
C
Organized
Self-disciplined
Neat and tidy
Disorganized
Responsible
Hard worker
A/N
Strong willed
Hardheaded
Stubborn
Gives in to others (rc)
Interested in new things
Disobedient
N
Feelings hurt
Easily upset
Afraid of lots
Whiny
Fearful
Sensitive to others
O
Thinking abilities
Good memory
Large vocabulary
Intelligent
Interested in new things
Asks questions
C
Organized
Neat and tidy
Careful
Responsible
Dependable
Self-disciplined
Figure 1
Comparison of Level 5 components from country-based samples.
both C/O and E contribute to a new “typical” A component. Finally,
at Level 5, the antagonism A/N component breaks down into separate antagonism and N components. In the Chinese sample, we again
see both a “typical” A and an antagonism A/N component represented. At Level 5, Chinese parents are not differentiating between C
and O. In addition, component 5, the antagonism A/N, is weakly
supported, with only six items loading substantially on this component. Variance accounted for at Levels 1–5 was 18.49%, 26.84%,
33.07%, 37.03%, and 39.96%, respectively.
14
Tackett, Slobodskaya, Mar, et al.
Greece. Level 2 of the personality hierarchy for Greek children
indicated two broad components reflecting E/O and antagonism
A/N. At Level 3, a C/O component breaks apart from both E/O and
A/N, which contribute roughly equivalent amounts of variance. At
Level 4, N weakly breaks off from E, as evidenced by a small positive
correlation between these two components. This N component
also includes typical Agreeableness items such as “considerate” and
“thoughtful” alongside N items reflecting sensitivity and fearfulness.
Finally, at Level 5, C/O separates into distinguishable C and O
components. Variance accounted for at Levels 1–5 was 20.07%,
30.16%, 36.06%, 40.90%, and 43.18%, respectively.
Russia. Level 2 of the personality hierarchy for Russian children
indicated two broad components reflecting E/O and A/N. At Level 3,
A/N breaks into A/C and an antagonism-heavy A/N. At Level 4, A/C
breaks into C/O and a typical A component. Finally, at Level 5, C/O
breaks apart into distinguishable C and O components. For Russian
children, we again see an antagonism A/N rather than a typical N
component. Variance accounted for at Levels 1–5 was 20.73%,
29.06%, 34.75%, 38.55%, and 41.05%, respectively.
United States. Level 2 of the personality hierarchy for American
children indicated two broad components reflecting O/E and A/N.
At Level 3, both components contribute to a C/O component, with
Level 2 O/E contributing more variance. At Level 4, the E and A/N
components contribute to an N component. Finally, at Level 5, C/O
separates into distinguishable C and O components. At Level 5 in
American children, we do see a typical N component as well as
the antagonism A component; however, no typical A component
emerged. Variance accounted for at Levels 1–5 was 23.62%, 32.60%,
37.52%, 41.76%, and 43.99%, respectively.
Procrustes rotation. In order to provide empirical criteria for
evaluating the similarity of factor structures across countries, Procrustes rotation was employed to compare the Level 5 structures
across groups. As the ICID was initially developed in a U.S. sample,
the U.S. sample was selected as the target factor structure, each of
the other four countries was rotated to this target structure for those
factors appearing across both groups, and factor congruence coefficients were calculated (see Table 2). Following recommendations by
0.98
0.97
0.97
0.95
0.95
0.97
0.99
0.98
0.98
Comp. 2: C/A
Comp 1: E
0.99
0.99
0.98
0.97
0.90
0.95
0.94
Comp. 2: A/N
0.97
0.93
0.95
0.96
Comp. 1: E
0.98
0.93
0.95
0.98
0.94
0.92
Comp. 3: A/N
0.97
0.89
0.94
0.95
Comp. 3: O
0.98
0.88
0.89
0.95
0.87
0.86
Comp. 4: O
*
0.87
0.93
*
Comp. 4: N
*
0.81
*
0.85
0.25
*
Comp. 5: N
0.90
*
0.89
0.88
Comp. 5: C
Note. E = Extraversion; A = Agreeableness; N = Neuroticism; O = Openness to Experience; C = Conscientiousness. * indicates that a
component with directly comparable content was not extracted at Level 5.
Target: 3–5-year-olds
6–8-year-olds
9–11-year-olds
12–14-year-olds
Target: 6–8-year-olds
9–11-year-olds
12–14-year-olds
Target: 9–11-year-olds
12–14-year-olds
Canada
China
Greece
Russia
Target: U.S. Sample
Table 2
Factor Congruence Coefficients From Orthogonal Procrustes Rotation of Level 5 Components
Child Personality Structure
15
16
Tackett, Slobodskaya, Mar, et al.
Fischer and Fontaine (2011), values lower than 0.85 are interpreted
as indicating factor incongruence, with desirable congruence coefficients falling at 0.95 or higher. Comparisons revealed adequate to
good average factor congruence coefficients for all samples: Canada
(0.95), China (0.90), Greece (0.93), and Russia (0.93). When a
directly comparable component was not extracted in both countries,
no congruence coefficient was available. For example, the fourth
component extracted in the U.S. sample was primarily defined by
items reflecting N, whereas a similar component was not extracted
for Canadian or Russian children.
Personality Hierarchy Across Development
Ages 3–5. Level 2 of the personality hierarchy for 3–5-year-old
children indicated two broad components reflecting characteristics of
O/E and A/N, respectively (see Figure 2). The first component broke
off to form separate components reflecting C and E at Level 3. At
Level 4, C breaks down into C and O. Finally, at Level 5, both E and
A/N contribute to a separate N component. Variance accounted for
at Levels 1–5 was 20.11%, 27.94%, 33.27%, 36.90%, and 39.74%,
respectively.
Ages 6–8. Level 2 of the personality hierarchy for 6–8-year-old
children similarly indicated two broad components reflecting E/O
and A/N, respectively (see Figure 3). At Level 3, both components
contribute to a separate C component. At Level 4, the C component
breaks into C/O and a new “typical” A component, with C items also
loading on this component. The transition from Level 4 to Level 5
is straightforward for E and O, with a more complicated pattern
emerging for the remaining components. A face-valid C component
breaks off of A at Level 4. Both A/N and A at Level 4 contribute to
an antagonistic A/N component and an N component at Level 5.
Variance accounted for at Levels 1–5 was 19.85%, 28.56%, 34.17%,
38.23%, and 40.96%, respectively.
Ages 9–11. Level 2 of the personality hierarchy for 9–11-year-old
children was again consistent with earlier ages, in that two broad
components reflecting E/O and A/N emerged (see Figure 4). At Level
3, an O/C component breaks apart from both E/O and A/N. At Level
4, N weakly breaks off from both E and A/N. Finally, at Level 5, O
Figure 2
Hierarchical structure of childhood personality from caregiver ratings of children ages 3–5 (n = 908).
Child Personality Structure
17
Figure 3
Hierarchical structure of childhood personality from caregiver ratings of children ages 6–8 (n = 874).
18
Tackett, Slobodskaya, Mar, et al.
Figure 4
Hierarchical structure of childhood personality from caregiver ratings of children ages 9–11 (n = 1,302).
Child Personality Structure
19
20
Tackett, Slobodskaya, Mar, et al.
and C split apart, with Level 4 E contributing some variance to O as
well. Variance accounted for at Levels 1–5 was 22.28%, 31.37%,
36.42%, 40.85%, and 43.34%, respectively.
Ages 12–14. Level 2 of the personality hierarchy for 12–14-yearolds also demonstrated evidence for E/O and A/N, achieving robust
replication across childhood development at this level (see Figure 5).
At Level 3, C breaks off from both E/O and A/N, with Level 2 E/O
contributing more variance. At Level 4, C breaks into C and A. At
Level 5, E, A, C, and A/N are retained, whereas both C and E at
Level 4 contribute to a new O component. Variance accounted for
at Levels 1–5 was 24.73%, 33.15%, 38.47%, 43.18%, and 45.37%,
respectively.
Procrustes rotation. Procrustes rotation was again employed to
compare the Level 5 structures across age groups. As there was no
clear target sample, all possible pair-wise comparisons were performed across age groups (see Table 2). Using the 3–5-year-old
sample as the target, average congruence coefficients were 6–8 (0.95),
9–11 (0.80), and 12–14 (0.93). Using the 6–8-year-old sample as the
target, average congruence coefficients were 9–11 (0.91) and 12–14
(0.94). Finally, comparing the 12–14-year-old sample to the target
sample of 9–11-year-olds produced an average congruence coefficient
of 0.98. When a directly comparable component was not extracted in
both age groups, no congruence coefficient was available.
DISCUSSION
This study provides the first examination of the hierarchical structure
of childhood personality across five different countries and four age
groups. Overall, more similarities were found than differences across
both country and age. For all countries and age groups, the general
traits extracted at Levels 2–5 map onto established patterns for
children and adults (Markon et al., 2005; Tackett et al., 2008). That
is, in all analyses, Level 2 typically reflected a difference in approachrelated characteristics versus avoidance and regulatory characteristics (DeYoung, 2006; Digman, 1997; Read et al., 2010). Level 3
reflected components resembling effortful control, positive emotionality, and negative emotionality (Rothbart et al., 2001). Level 4
Figure 5
Hierarchical structure of childhood personality from caregiver ratings of children ages 12–14 (n = 649).
Child Personality Structure
21
22
Tackett, Slobodskaya, Mar, et al.
generally resembled the FFM without a separate O factor, whereas
at Level 5 all five factors were typically represented. Some deviations
from previous findings were also observed, as discussed below.
Childhood Personality Structure Across Cultures
The countries that would typically be characterized as individualistic
(i.e., Canada and the United States) were largely similar to one
another and more similar to previous findings than the collectivistic
countries (i.e., China and Russia). Although Greece is typically
ranked between largely individualistic and collectivistic cultures, it is
ranked more similarly to collectivistic cultures (Hofstede, 2001; Suh
et al., 1998). Yet, in these findings, the hierarchical structure of
Greek children appeared to more closely resemble that of Canadian
and American children. These overall similarities are also reflected
when examining factor congruence coefficients between countries at
Level 5 of the hierarchy, which support good overall replication of
Level 5 traits across samples. In particular, E, A/N, and O showed
the most robust replication across countries, with less robust replication for pure N and C traits.
One interesting finding is that the pattern of covariation for O was
identical in all three individualistic countries. Specifically, O characteristics covaried with E at Level 2 but shifted to C for Levels 3 and
4 in Canada, Greece, and the United States. These findings were not
expected based on the adult literature, but derive some support from
other childhood personality findings. C and O in childhood have
been linked to academic achievement in several studies, for example
(Barbaranelli et al., 2003; John et al., 1994; Mervielde, Buyst, & De
Fruyt, 1995), and previous research on childhood personality has
found stronger relationships between C and O than would be
expected in adult samples (De Pauw et al., 2009; Goldberg, 2001).
The predominance of intellect features for O at early ages may be the
factor that explains this stronger covariation (Gjerde & Cardilla,
2009). One study of adolescents (ages 12–17) found strong evidence
for a factor consisting of both Openness to Experience and intellect,
although relations between these scales and an adult O measure were
stronger with O than intellect, suggesting that adult personality measures may not cover the intellect domain as thoroughly (De Fruyt
et al., 2000). Specifically, such results suggest that bottom-up measures of childhood personality may find intellect features to play a
Child Personality Structure
23
more primary role in personality structure than they do in adult
measures. The relation between childhood measures of O and adult
personality are complex. Some findings have suggested differential
relations depending on gender, such that child O predicts adult O for
females, whereas in males child O is a stronger predictor of adult C
(Gjerde & Cardilla, 2009). Thus, the role of early intellect, imagination, and curiosity may hold different predictive power for males and
females. Further, these aspects may be stronger representations of O
in early life than in adulthood, a topic that should be studied in
future research.
Shifting our focus to the other countries, the pattern seen for the
children in China (ranked most collectivistic) is particularly interesting. This is the only country where separate components for C and O
did not emerge at Level 5; rather, they appear to represent a unified
factor even at this level. This is also reflected when comparing this
component to the O component in the U.S. sample, with a factor
congruence coefficient of 0.89. This is somewhat consistent with
previous cross-cultural research that has often failed to find a robust
O analog in non-Western samples (e.g., Cheung et al., 2001). Similarly, in the Russian sample, the O characteristics did not play a
particularly prominent role at higher levels of the hierarchy, with no
items of O appearing as substantial loadings on the components at
Level 3. By Level 5, however, the factor congruence coefficient suggests that a largely replicated O factor did emerge in the Russian
sample.
Childhood Personality Structure Across Ages
Turning to the results across age groups, the hypothesis that Levels
2 and 3 would be robust across age was supported. Somewhat surprisingly, however, was the finding that Levels 4 and 5 were largely
consistent across age as well. Even in the youngest age group, the
four- and five-factor structures look very similar to those found in
adult samples. Some interesting inconsistencies emerged as well.
Examination of factor congruence coefficients between age groups
for Level 5 components suggest that a pure N component showed
the most difficulty in replication across ages. This may be due, in
part, to the difficulty in measuring more internal aspects of N (e.g.,
sadness, anxiety, insecurity) in children when relying on informant
reports.
24
Tackett, Slobodskaya, Mar, et al.
We also observed a shifting pattern for C, such that in the younger
age groups (3–5 years and 6–8 years), C is predominantly loading
with A characteristics reflecting the previously described “agreeable
compliance” (DePauw & Mervielde, 2010; Tackett et al., 2008). With
the older age groups (9–11 years and 12–14 years), we see a much
cleaner differentiation of C and A characteristics such that C and A
items are not covarying as tightly at higher levels of the hierarchy.
This is consistent with the broad disconstraint factor that is primarily
represented in temperament models and thought to subsume aspects
of both C and A (Rothbart et al., 2001). These findings suggest that
A and C do tightly covary at younger ages, which may represent a
true lack of differentiation between A and C in early childhood or
may be a result of measurement limitations (or both). For example,
many aspects of C may not be salient until the child is fully immersed
in the academic setting that primary school provides. It may be that
early childhood environments restrict opportunities for such traits to
manifest. Nonetheless, by Level 5, the emergent C trait (whether
primarily C or still including some A items as well) showed robust
replicability across age groups, suggesting good continuity in measurement of this trait across time. Future work using multiple informants and methods of measurement should aim to disentangle
measurement bias from true variance in these traits.
Contributions to Conceptualizations of Childhood Personality
Similar to research findings with adults, this study supports a robust
hierarchical trait structure of child personality across age groups and
cultures. This vast replicability, with more similarities than differences across groups, provides further support for hierarchy as an
inherent and important aspect of personality structure (Markon,
2009). Establishing trait hierarchy allows for a number of theoretical
and practical implications. The robust nature of superordinate traits
is easily integrated with major causal and explanatory theories of
trait structure. For example, rich explanatory theories have been
developed for the “Big Two” (seen here at the second level of
the hierarchy), connecting these broad traits to neurobiological
(Markon, 2009) and motivational (Read et al., 2010) systems that
may serve to link causal factors, mediating processes, traits, and
behavior. Importantly, these findings support the search for such
explanatory pathways from early childhood through adulthood. In
Child Personality Structure
25
addition, these results suggest that future investigations on underlying explanatory processes may be best served by extracting even
higher levels of superordinate traits from existing measures in order
to provide continuity and to ease interpretation of results across
studies.
One fairly robust finding that deviates from previous research
with adults is the salient role of antagonism in these models. Across
ages, we see that emergence of typical A (e.g., marked by empathy,
compassion, and modesty) and typical N (e.g., marked by sadness
and depression) traits is hard to find in these childhood data. Instead,
the antagonistic features associated with both A and N emerge early
(i.e., at Level 2) and remain a very salient trait that parents use to
differentiate their children. This is consistent with the extensive literature on the “difficult temperament” construct, which sometimes
incorporates both negative affectivity and behavioral resistance or
oppositionality (Tackett, 2006). A “typical” N emerges by Level 5 for
ages 3–5, 6–8, and 9–11, yet a “typical” A does not. The A at Level
5 for these ages is still largely reflecting antagonism. The pattern
reverses in the oldest group, the 12–14-year-olds, where we finally see
a “typical” A at Level 5, although negative affectivity is still primarily defined by antagonism rather than sad or anxious mood.
These strong links between A and N are not new. The adult
temperament model developed by Evans and Rothbart (2007) distinguished between aggressive and nonaggressive negative affect.
While nonaggressive negative affect correlated primarily with FFM
N, aggressive negative affect correlated substantially with both N
and A (inversely) in a college student sample. In recent reviews, the
lower order trait of anger-irritability was noted to load differentially
at the higher order level depending on the model of childhood
personality or temperament, with some temperament measures
more commonly assigning it to Neuroticism/negative emotionality,
whereas personality measures more often assigned it to Agreeableness (Caspi & Shiner, 2006; De Pauw & Mervielde, 2010). In addition, childhood N is likely the most difficult trait to measure using
informant reports, as it is less directly observable than the other traits
(Hampson & Goldberg, 2006). Indeed, informant reports of childhood N show the weakest prediction of later self-reported personality relative to other traits (Hampson & Goldberg, 2006; Tackett
et al., 2008). These findings suggest that antagonistic aspects of N
may be more easily identifiable at earlier ages than other facets of this
26
Tackett, Slobodskaya, Mar, et al.
domain. As children move into adolescence, negative affectivity
increases (Carstensen, Pasupathi, Mayr, & Nesselroade, 2000), but
the parental exposure to this is likely to be primarily in the form of
antagonism. Youth are beginning to establish autonomy from their
parents at this developmental stage and may be less likely to share
their inner feelings and experiences related to depression and anxiety.
This study investigated the hierarchical structure of childhood
personality using a large pool of items from an existing measure that
maps onto a five-factor structure (Halverson et al., 2003). Recent
researchers examining the joint structure of temperament and personality have suggested the possibility of a six-factor model to incorporate both personality and temperament domains (De Pauw et al.,
2009; De Pauw & Mervielde, 2010). Specifically, De Pauw and
Mervielde (2010) recommend a distinct Activity domain at the
higher order level that is separate from E and is not present in adult
personality models. Other taxonomic approaches define Activity
as a subcomponent of E (Caspi et al., 2005). Importantly, Activity
showed poor congruence across samples of Russian, Slovenian, and
U.S. children (Knyazev et al., 2008), so more research investigating
the cross-cultural generalizability of this trait is needed. Given the
incipient nature of this research area, future investigations more
broadly tapping the joint domains covered by temperament and
personality models will be necessary to further understand the structure of childhood personality across ages. Temperament research has
much to offer investigations of childhood personality, including an
extensive literature with attention to developmental processes and
the biological substrates of traits (Rothbart et al., 2000). Future
work in this area should continue to move toward further integration
of temperament and personality at early ages.
One important contribution this study makes is an examination of
personality across countries and ages from the perspective of structure and item-level covariation rather than mean trait levels. Much
of the existing adult work examining cross-cultural differences
in personality has focused on mean-level differences, an approach
that involves a number of limitations regarding validity (Heine &
Buchtel, 2009). Examination of covariation patterns allows a look at
cross-country differences from a different angle, potentially avoiding
some of the previous pitfalls. For example, mean-level differences
may be more influenced by biases such as cultural reference groups,
whereas item-level covariation is examining relative patterns rather
Child Personality Structure
27
than absolute values; thus, comparisons across groups may be less
influenced by such reporting biases.
Limitations
One limitation of this study is the reliance on parental report, primarily from the mother. Development of valid and reliable selfreport measures of childhood personality has been slow, due to the
inherent social-cognitive and intellectual limitations in early stages
of development (Markey, Markey, Tinsley, & Ericksen, 2002).
Researchers have begun identifying potentially fruitful methods of
assessing child personality via self-report (e.g., Brown, Mangelsdorf,
Agathen, & Ho, 2008; Eder, 1990; Markey et al., 2002; Measelle,
John, Ablow, Cowan, & Cowan, 2005), although many are quite
resource-intensive. Even at relatively late stages of childhood development (10–12 years of age), self-reports are potentially problematic,
with lower reliabilities than parent report (Markey et al., 2002). In
addition, acquiescent responding is more prevalent in later childhood and decreases substantially from age 10 to late adolescence
(Soto et al., 2008).
Parent reports have certain advantages regarding their psychometric properties, and it is important to note that they appear
to provide converging evidence with child self-reports. Previous
research examined factor congruence coefficients for parent and selfreport of early adolescents on the ICID lower order scales and found
a highly comparable structure for both parent and self-report, with
the exception of the Activity scale (Knyazev et al., 2008). Further,
parent and self-report are correlated in early adolescence for all traits
(Markey et al., 2002). These findings suggest that parental reports of
childhood personality do contain valid information about early individual differences (Rothbart & Bates, 2006), although such work
would still be strengthened by the future use of multiple informants
and methods. It is also likely that parent and teacher informants are
better assessors of certain childhood personality traits than others.
For example, both parent and teacher ratings of C and E showed the
highest prediction of the analogous traits self-reported in later life
(Hampson & Goldberg, 2006; Tackett et al., 2008) in comparison to
other traits. Such findings could demonstrate higher stability of C
and E over time, but they could also reflect more accurate measurement of these traits when relying on parent and teacher ratings for
28
Tackett, Slobodskaya, Mar, et al.
childhood personality (see also Measelle et al., 2005). Thus, a more
intensive and detailed approach to measuring childhood personality
should involve a combination of informants (e.g., self, mother,
father, teacher, peer) to more accurately assess the full range of traits.
Another limitation of the current study is the absence of a refined
measure of culture and the ability to examine the dimensions of age
and culture independently of one another. The countries examined
here represent both individualistic and collectivistic cultures, but
culture is more complex than simply country of origin (Kotelnikova
& Tackett, 2010). Furthermore, the country-based samples in the
present article were collected independently, restricting the availability of standardized demographic information (e.g., ethnicity)
across samples. More careful attention to the role that culture plays
in childhood personality and personality development may offer
important insight into what may be the most important environmental influence on personality traits (Super & Harkness, 2002). Little
cross-cultural work on childhood personality has been conducted,
and this remains an important and exciting area of future study. In
addition to examining cross-cultural differences in emergent personality, potential cultural differences in informants and methods must
also be considered. One study investigated the values that mothers
from Greece, Taiwan, and the United States wished to instill in their
children and found some interesting differences (Tamis-LeMonda,
Wang, Koutsouvanou, & Albright, 2002). Greek mothers emphasized values such as honesty and spirituality, Taiwanese mothers
emphasized values such as good manners and obedience, and American mothers emphasized values such as assertiveness and independence as well as values such as sociability and compassion. Similarly,
parental perceptions of the same trait may yield different consequences (e.g., reward vs. punishment) across cultures (e.g., Chen
et al., 1998), which may lead to differential reporting of mean levels
across groups (and differential expression of these traits across development). It is possible that cultural differences such as these may
influence parent reports of childhood personality across cultures.
CONCLUSION
This study provides the strongest evidence to date regarding the
universality of hierarchical personality structure in childhood by
Child Personality Structure
29
incorporating multiple cross-country and cross-age comparisons
(Bornstein, 2002). In a large sample of 3,751 children from Canada,
China, Greece, Russia, and the United States, robust similarities for
childhood personality structure emerged. Differences emerged as
well, both across development and across countries, with countries
typically classified as individualistic (i.e., Canada and the United
States) appearing more similar to one another than to the countries
typically classified as collectivistic (i.e., China and Russia). In these
samples, the Greek children appeared more similar to the individualistic than the collectivistic samples. Taken together, these results
help explain previous differences in temperament and childhood personality research, including the strong emergence of a robust threefactor structure and the concept of “difficult temperament,” which
incorporates both negative emotionality and (dis)agreeableness in
childhood. Further, support was found for an FFM from early childhood through early adolescence that appears largely analogous to
the established FFM in adult populations. These findings will help to
integrate child temperament and personality research with research
on adults, resulting in a useful refinement of theoretical conceptions
and opening up future avenues for child personality research.
REFERENCES
Ahadi, S. A., Rothbart, M. K., & Ye, R. (1993). Children’s temperament in the US
and China: Similarities and differences. European Journal of Personality, 7,
359–377.
Allik, J., Laidra, K., Realo, A., & Pullmann, H. (2004). Personality development
from 12 to 18 years of age: Changes in mean levels and structure of traits.
European Journal of Personality, 18, 445–462.
Barbaranelli, C., Caprara, G. V., Rabasca, A., & Pastorelli, C. (2003). A questionnaire for measuring the Big Five in late childhood. Personality and Individual Differences, 34, 645–664.
Bornstein, M. H. (2002). Toward a multiculture, multiage, multimethod science.
Human Development, 45, 257–263.
Brown, G. L., Mangelsdorf, S. C., Agathen, J. M., & Ho, M. H. (2008). Young
children’s psychological selves: Convergence with maternal reports of child
personality. Social Development, 17, 161–182.
Carstensen, L. L., Pasupathi, M., Mayr, U., & Nesselroade, J. R. (2000). Emotional experience in everyday life across the adult life span. Journal of Personality and Social Psychology, 79, 644–655.
Caspi, A., Roberts, B. W., & Shiner, R. L. (2005). Personality development:
Stability and change. Annual Review of Psychology, 56, 453–484.
30
Tackett, Slobodskaya, Mar, et al.
Caspi, A., & Shiner, R. L. (2006). Personality development. In N. Eisenberg (Vol.
Ed.) and W. Damon & R. M. Lerner (Eds.), Handbook of child psychology:
Social, emotional, and personality development (6th ed., Vol. 3, pp. 300–364).
Hoboken, NJ: Wiley.
Chen, X., Hastings, P. D., Rubin, K. H., Chen, H., Cen, G., & Stewart, S. L.
(1998). Child-rearing attitudes and behavioral inhibition in Chinese and
Canadian toddlers: A cross-cultural study. Developmental Psychology, 34, 677–
686.
Cheung, F. M., Leung, K., Zhang, J. X., Sun, H. F., Gan, Y. Q., Song, W. Z.,
et al. (2001). Indigenous Chinese personality constructs. Journal of CrossCultural Psychology, 32, 407–433.
Church, A. T. (2001). Personality measurement in cross-cultural perspective.
Journal of Personality, 69, 979–1006.
Church, A. T., & Lonner, W. J. (1998). The cross-cultural perspective in the study
of personality: Rationale and current research. Journal of Cross-Cultural Psychology, 29, 32–62.
De Fruyt, F., Bartels, M., Van Leeuwen, K. G., De Clercq, B., Decuyper, M., &
Mervielde, I. (2006). Five types of personality continuity in childhood and
adolescence. Journal of Personality and Social Psychology, 91, 538–552.
De Fruyt, F., Mervielde, I., Hoekstra, H. A., & Rolland, J. P. (2000). Assessing
adolescents’ personality with the NEO PI-R. Assessment, 7, 329–345.
De Pauw, S. S., & Mervielde, I. (2010). Temperament, personality and developmental psychopathology: A review based on the conceptual dimensions underlying childhood traits. Child Psychiatry and Human Development, 41, 313–329.
De Pauw, S. S., Mervielde, I., & Van Leeuwen, K. G. (2009). How are traits
related to problem behavior in preschoolers? Similarities and contrasts between
temperament and personality. Journal of Abnormal Child Psychology, 37, 309–
325.
De Young, C. G. (2006). Higher-order factors of the Big Five in a multi-informant
sample. Journal of Personality and Social Psychology, 91, 1138–1151.
Digman, J. M. (1997). Higher-order factors of the Big Five. Journal of Personality
and Social Psychology, 73, 1246–1256.
Digman, J. M., & Shmelyov, A. G. (1996). The structure of temperament and
personality in Russian children. Journal of Personality and Social Psychology,
71, 341–351.
Eder, R. A. (1990). Uncovering young children’s psychological selves: Individual
and developmental differences. Child Development, 61, 849–863.
Evans, D. E., & Rothbart, M. K. (2007). Developing a model for adult temperament. Journal of Research in Personality, 41, 868–888.
Evans, D. E., & Rothbart, M. K. (2009). A two-factor model of temperament.
Personality and Individual Differences, 47, 565–570.
Fischer, R., & Fontaine, J. R. J. (2011). Methods for investigating structural
equivalence. In D. Matsumoto & F. J. R. van de Vijver (Eds.), Cross-cultural
research methods in psychology (pp. 179–215). New York: Cambridge University Press.
Gartstein, M. A., Peleg, Y., Young, B. N., & Slobodskaya, H. R. (2009). Infant
temperament in Russia, United States of America, and Israel: Differences and
Child Personality Structure
31
similarities between Russian-speaking families. Child Psychiatry and Human
Development, 40, 241–256.
Gjerde, P. F., & Cardilla, K. (2009). Developmental implications of openness to
experience in preschool children: Gender differences in young adulthood.
Developmental Psychology, 45, 1455–1464.
Goldberg, L. R. (2001). Analyses of Digman’s child-personality data: Derivation
of Big-Five factor scores from each of six samples. Journal of Personality, 69,
710–743.
Goldberg, L. R. (2006). Doing it all bass-ackwards: The development of hierarchical factor structures from the top down. Journal of Research in Personality,
40, 347–358.
Halverson, C. F., Havill, V. L., Deal, J., Baker, S. R., Victor, J. B., Pavlopoulos,
V., et al. (2003). Personality structure as derived from parental ratings of free
descriptions of children: The inventory of child individual differences. Journal
of Personality, 71, 995–1026.
Hampson, S. E. (2008). Mechanisms by which childhood personality traits
influence adult well-being. Current Directions in Psychological Science, 17,
264–268.
Hampson, S. E., & Goldberg, L. R. (2006). A first large cohort study of personality trait stability over the 40 years between elementary school and midlife.
Journal of Personality and Social Psychology, 91, 763–779.
Heine, S. J., & Buchtel, E. E. (2009). Personality: The universal and the culturally
specific. Annual Review of Psychology, 60, 369–394.
Hofstede, G. (2001). Culture’s consequences: Comparing values, behaviors, institutions and organizations across nations (2nd ed.). Beverly Hills, CA: Sage.
John, O. P., Caspi, A., Robins, R. W., Moffitt, T. E., & Stouthamer-Loeber, M.
(1994). The “Little Five”: Exploring the nomological network of the five-factor
model of personality in adolescent boys. Child Development, 65, 160–178.
John, O. P., Naumann, L. P., & Soto, C. J. (2008). Paradigm shift to the integrative Big Five trait taxonomy. In O. P. John, R. W. Robins, & L. A. Pervin
(Eds.), Handbook of personality: Theory and research (3rd ed., pp. 114–158).
New York: Guilford Press.
Knyazev, G. G., & Slobodskaya, H. R. (2005). Five factor personality structure in
children and adolescents (based on parent and self-reports). Psychological
Journal, 26, 59–67. (In Russian).
Knyazev, G. G., Zupancic, M., & Slobodskaya, H. R. (2008). Child personality in Slovenia and Russia. Journal of Cross-Cultural Psychology, 39, 317–
334.
Kohnstamm, G. A., Halverson, C. F., Jr., Mervielde, I., & Havill, V. L. (Eds.).
(1998). Parental descriptions of child personality: Developmental antecedents of
the Big Five? Hillsdale, NJ: Erlbaum.
Kotelnikova, Y., & Tackett, J. L. (2010). Values as predictors of psychopathology
in a multicultural Canadian sample. Poster presented at the 24th annual meeting
of the Society for Research on Psychopathology, Seattle, WA.
Lee, K., & Ashton, M. C. (2008). The HEXACO personality factors in the
indigenous personality lexicons of English and 11 other languages. Journal of
Personality, 76, 1001–1053.
32
Tackett, Slobodskaya, Mar, et al.
Markey, P. M., Markey, C. N., Tinsley, B. J., & Ericksen, A. J. (2002). A preliminary validation of preadolescents’ self-reports using the five-factor model of
personality. Journal of Research in Personality, 36, 173–181.
Markon, K. E. (2009). Hierarchies in the structure of personality traits. Social and
Personality Psychology Compass, 3, 812–826.
Markon, K. E., Krueger, R. F., & Watson, D. (2005). Delineating the structure of
normal and abnormal personality: An integrative hierarchical approach.
Journal of Personality and Social Psychology, 88, 139–157.
Martel, M. M., Nigg, J. T., & Lucas, R. E. (2008). Trait mechanisms in youth with
and without attention-deficit/hyperactivity disorder. Journal of Research in
Personality, 42, 895–913.
McCrae, R. R., Costa, P. T., Jr., Terracciano, A., Parker, W. D., Mills, C. J., De
Fruyt, F., et al. (2002). Personality trait development from age 12 to age 18:
Longitudinal, cross-sectional, and cross-cultural analyses. Journal of Personality and Social Psychology, 83, 1456–1468.
Measelle, J. R., John, O. P., Ablow, J. C., Cowan, P. A., & Cowan, C. P. (2005).
Can children provide coherent, stable, and valid self-reports on the Big Five
dimensions? A longitudinal study from ages 5 to 7. Journal of Personality and
Social Psychology, 89, 90–106.
Mervielde, I., Buyst, V., & De Fruyt, F. (1995). The validity of the Big Five as a
model for teachers’ ratings of individual differences in children aged 4 to 12.
Personality and Individual Differences, 18, 525–534.
Mervielde, I., & De Fruyt, F. (1999). Construction of the Hierarchical Personality
Inventory for Children (HiPIC). In I. Mervielde, I. Deary, F. De Fruyt, & F.
Ostendorf (Eds.), Personality psychology in Europe: Proceedings of the Eighth
European Conference on personality psychology (pp. 107–127). Tilburg, The
Netherlands: Tilburg University Press.
Read, S. J., Monroe, B. M., Brownstein, A. L., Yang, Y., Chopra, G., & Miller,
L. C. (2010). A neural network model of the structure and dynamics of human
personality. Psychological Review, 117, 61–92.
Roberts, B. W., & DelVecchio, W. F. (2000). The rank-order consistency of
personality traits from childhood to old age: A quantitative review of longitudinal studies. Psychological Bulletin, 126, 3–25.
Roberts, B. W., Walton, K. E., & Viechtbauer, W. (2006). Patterns of mean-level
change in personality traits across the life course: A meta-analysis of longitudinal studies. Psychological Bulletin, 132, 1–25.
Rothbart, M. K., Ahadi, S. A., & Evans, D. E. (2000). Temperament and personality: Origins and outcomes. Journal of Personality and Social Psychology, 78,
122–135.
Rothbart, M. K., Ahadi, S. A., Hershey, K., & Fisher, P. (2001). Investigations of
temperament at 3 to 7 years: The Children’s Behavior Questionnaire. Child
Development, 72, 1394–1408.
Rothbart, M. K., & Bates, J. E. (2006). Temperament. In N. Eisenberg (Vol. Ed.)
and W. Damon & R. M. Lerner (Eds.), Handbook of child psychology: Social,
emotional, and personality development (6th ed., Vol. 3, pp. 99–166). Hoboken,
NJ: Wiley.
Child Personality Structure
33
Rubin, K. H., Hemphill, S. A., Chen, X., Hastings, P., Sanson, A., Lo Coco, A.,
et al. (2006). A cross-cultural study of behavioral inhibition in toddlers: EastWest-North-South. International Journal of Behavioral Development, 30, 219–
226.
Schwartz, S. H. (1992). Universals in the content and structure of values: Theory
and empirical tests in 20 countries. In M. Zanna (Ed.), Advances in experimental social psychology (Vol. 25, pp. 1–65). New York: Academic Press.
Shiner, R., & Caspi, A. (2003). Personality differences in childhood and adolescence: Measurement, development, and consequences. Journal of Child
Psychology and Psychiatry, 44, 2–32.
Soto, C. J., John, O. P., Gosling, S. D., & Potter, J. (2008). The developmental
psychometrics of Big Five self-reports: Acquiescence, factor structure, coherence, and differentiation from ages 10 to 20. Journal of Personality and Social
Psychology, 94, 718–737.
Suh, M., Diener, E., Oishi, S., & Triandis, H. C. (1998). The shifting basis of life
satisfaction judgments across cultures: Emotions versus norms. Journal of
Personality and Social Psychology, 74, 482–493.
Super, C. M., & Harkness, S. (2002). Culture structures the environment for
development. Human Development, 45, 270–274.
Tackett, J. L. (2006). Evaluating models of the personality-psychopathology relationship in children and adolescents. Clinical Psychology Review, 26, 584–599.
Tackett, J. L., Krueger, R. F., Iacono, W. G., & McGue, M. (2008). Personality in
middle childhood: A hierarchical structure and longitudinal connections with
personality in late adolescence. Journal of Research in Personality, 42, 1456–
1462.
Tamis-LeMonda, C. S., Wang, S., Koutsouvanou, E., & Albright, M. (2002).
Childrearing values in Greece, Taiwan, and the United States. Parenting:
Science and Practice, 2, 185–208.
Tellegen, A., & Waller, N. G. (1992). Exploring personality through test construction: Development of the Multi-dimensional Personality Questionnaire
(MPQ). Unpublished manuscript, Department of Psychology, University of
Minnesota.
Windle, M., Iwawaki, S., & Lerner, R. M. (1988). Cross-cultural comparability of
temperament among Japanese and American preschool children. International
Journal of Psychology, 23, 547–567.